Yahoo Finance·2026-02-27 10:51Group 1: AI Infrastructure Spending - AI infrastructure spending is expected to rise significantly in 2026, with the five largest U.S. hyperscalers and cloud computing companies projected to spend between $660 billion and $690 billion on AI infrastructure, compared to $380 billion last year [1] - Pure-play AI companies and ambitious infrastructure projects will contribute to this spending, highlighting the importance of hardware providers like Micron Technology and Jabil [2] Group 2: Micron Technology's Role - Micron Technology's DRAM chips are essential for AI data centers, particularly high-bandwidth memory (HBM), which enables AI accelerator chips to efficiently access large datasets for training and inference [4] - AI servers require six times the amount of DRAM compared to standard servers, a prediction made by Micron in 2019 that has proven accurate as demand for DRAM chips has surged [5] - Nvidia's new AI processors are expected to utilize nearly 300 GB of HBM, up from 200 GB in previous models, contributing to a significant shortage of memory chips and driving prices up by 75% from December 2025 to January 2026 [6] - The trend of rising memory prices is anticipated to continue due to ongoing investments in AI infrastructure and supply constraints, leading to positive earnings growth prospects for Micron Technology [7]